You are on page 1of 24

talk given at the School

Understanding our Universe at the close of 20th Century
25 April 2000 - 6 May 2000
Cargese, France

arXiv:astro-ph/0010634v1 31 Oct 2000

Supernovae and Cosmology
Monique Signore⋆, Denis Puy†,‡
⋆ Laboratoire

de Radioastronomie, Observatoire de Paris (France)
Scherrer Institut, Villigen (Switzerland)
‡ Institute of Theoretical Physik, University of Zurich (Switzerland)
† Paul

These lecture notes intend to form a short pedagogical introduction to the use of
typical type Ia-Supernovae (hereafter SNIa) as standard candles to determine the
energy density of the universe. Problems of principle for taking SNIa as cosmological
probes are pointed out, and new attempts at solving them are indicated including
the empirical width-luminosity relation (WLR) and its possible explanations.
Finally, the observations of SNIa at high redshift carried out by two major teams
are briefly reviewed and their interpretation as evidence for an accelerating universe
is also rapidly discussed
Key words: Supernovae, Observational cosmology
PACS: 97.60.Bw, 98.80.Es



A supernova is a very powerful event: a star which suddenly brightens to about
109 -1010 L⊙ . There are two different types of supernovae (see Table 1): SNII
and SNIa, such as:
• SNII are core-collapse induced explosions of short-lived massive stars
(M⋆ > 8 M⊙ ) which produce more O and Mg relative to F e.
• SNIa are thermonuclear explosions of accreting white-dwarfs in close binaries which produce mostly F e and little O; the exact companions stars of


Elsevier Preprint

1 February 2008

white-dwarfs are generally not identified but must be relatively long-lived
There are also particular supernovae: SNIb and SNIa with most properties
of SNII except their spectra which have no H line (such as SNIa) but they
believed to be core-collapse supernovae (such as SNII).





Metal lines
deep 6180 ˚

P Cyg lines
Balmer series

∼ 4×109 L⊙
small dispersion
standard candles ?

∼ 109 L⊙
large dispersion



very weak


Radio emission

no detection

slow decay


all galaxies

spirals and irregulars



WD in binary systems

massive stars

Optical spectrum

Absolute luminosity
at max. light
Optical light curve
UV spectrum

Stellar population

For some time astronomers had focused on all types of supernovae to measure Hubble’s constant. Now, with the use of Hubble Space Telescope (HST)
observations of Cepheids in the host galaxies of these supernovae, there is a
consensus that Ho is in the range of 60-70 km s−1 Mpc−1 -see Branch (1998)
for a review.
But very recently, two observational groups working independently to use
SNIa as standard candles to measure distance versus redshift presented evidence that the Hubble expansion has been accelerating over time. In effect,
these two independent groups (Perlmutter et al. 1998, Schmidt et al. 1998)
had to develop not only a method of discovering SNIa at high redshift by
careful observations -with large format CCDs, large apertures- and scanning
of plates but they also had to determine if these SNIa could be used as standard candles. This recent and important progress has been possible because
an empirical relation (Phillips 1993) between the duration of the peak phase
of a SNIa’s light curve and its luminosity (broader is brighter) has been taken
into account by both teams.
This is precisely the work of these two groups that we are trying to present, in

because they are among the brightest optical explosive events in the universe. their evolutionthe hydrodynamical models for SNIa. in these lessons. In this scope. 1997). for more insight into the underlying physics of SNIa and their progenitors.are still uncertain. SNIa have been tentatively used as standard educational manner.the mass of the WD at ignition. ii) The sub-Chandrasekhar mass model.broader and brighterPhillips (1993). 3 . γ-rays from SNIas. the progenitor systems of SNIa -their nature. Cosmic Explosions (Holt & Zhang 2000). 2. 2 On type Ia supernovae There are spectroscopic and photometric indications that SNIa result from the thermonuclear explosions of accreting carbon/oxygen white dwarfs (hereafter WD). in section 2. 1990). However. we give a cosmological background by recalling what is the (ΩΛ . ΩM )-plane and the luminosity distance. In section 5.luminosity relation (WLR) which is also called the Phillips Relation. the physics of the nuclear burnings. we present the observations of SNIa at high redshift by the two major teams: the Supernovae Cosmology Project (SCP) and the High z supernova search team (HZT).38 M⊙ -which is near the Chandrasekhar mass and explodes as a SNIa.37 − 1. Livio (2000) and most of the papers of the following proceedings Supernovae (Petscheck. Thermonuclear Supernovae (Ruiz-Lapente et al. Nomoto et al. (1997). some models of their explosions. However variations of light curves and spectra among SNIa have recently been extensively studied. we discuss some aspects of their results and finally give a conclusion. observed optical spectra of SNIas. see the various excellent reviews of Woosley-Weaver (1986). in which an accreted layer of helium.1 On the progenitors of SNIa There are two classes of models proposed as progenitors of SNIa: i) The Chandrasekhar mass model in which a mass acreting C + O WD grows in mass up to the critical mass: MIa ∼ 1. Moreover. and because their light curves and their spectral evolution are relatively uniform. In section 3. Type Ia Supernovae and Cosmology (Niemeyer Truran 2000). in particular. This section briefly summarizes our knowledge of these SNIa properties and also their controversial issues. However. we summarize our current understanding of SNIas: their possible progenitors. the relation between the duration of the peak phase of their light curves and their luminosities . In section 4. observed optical light curves and finally the empirical width.

the merging of the C +O WD with a combined mass surpassing the Chandrasekhar mass Mch . 2. for these models.e.e. Moreover. the first important question is about ignition. a) a mass-accreting WD and a lobe-filling less massive red giant star (WD+RG system). Let us conclude this subsection by saying that the evolution of the progenitor is undoubtly the most uncertain part of a fully predictive model of a SNIa explosion. For the present binary systems which grow the WD mass to MIa there are two possible systems: a) a mass-accreting WD and a lobe-filling main sequences star (WD+MS system). The early time spectra of the majority of SNIa are in excellent agreement with the synthetic spectra of the Chandrasekhar mass models. two scenarii have been proposed: a) a double degenerate (DD) scenario.1.1. the accretion of hydrogen rich matter via mass transfer from a binary companion. The issue of DD versus SD scenario is still debated. The spectra of the sub-Chandrasekhar mass models are too blue to be comparable with the observations. However.2 On explosion models As already noted above. But let us only give some crude features about the progenitor evolution. b) a single degenerate (SD) scenario. 2. the accreted matter extends to form a common envelope. which can be formulated as: when and how does the burning start ? We have also 4 .2 Sub-Chandrasekhar mass models In the sub-Chandrasekhar mass model for SNIa. i.1 Chandrasekhar mass models For the evolution of accreting WD toward the Chandrasekhar mass model. i. a critical rate. 2. some theoretical modeling has indicated that the merging of WD lead to the accretion-induced collapse rather than a SNIa explosion.atop a C + O WD ignites off-center for a WD mass well below the Chandrasekhar mass. this difficulty has been recently overcome by a WD wind model. a WD explodes as a SNIa only when the rate of the mass accretion rate M˙ is in a certain narrow range. if M˙ > M˙ crit . Moreover.

6 M⊙ ).2.58 M⊙ of 56 Ni in the inner region is ejected with a velocity of 1-2 × 104 km s−1 . there has been great progress in applying concepts from terrestrial combustion physics to the SNIa problem.7 M⊙ < MW D < 1.e. ii) the synthesis of a large amount of 56 Ni (∼ 0. it has been shown that the outcome of carbon deflagration depends critically on its flame velocity 5 . for which: MW D ∼ Mch ∼ 1. ii) the ignition of He in a shell around a C + O core is followed by the explosive burning of the C + O core and the helium shell. 2.4 M⊙ .seen that there are eventually two candidates: i) the ignition of 12 C +12 C. In particular. let us only consider Chandrasekhar mass models as standard SNIa models: they provide a point of convergent evolution for various progenitor systems and therefore could offer a natural explanation of the assumed uniformity of display. The physics involved in these processes is very complex and relative to the physics of thermonuclear combustion. a subsonic deflagration wave propagates at an average speed of 1/5 of the sound speed from the center to the expanding outer layers. : i) an explosion energy of about 1051 ergs. Here. In particular. In this model. 2.4 M⊙ . the physics of turbulent combustion and the possible spontaneous transition to detonation are probably the most important and least tractable effects. this is the case of sub-Chandrasekhar mass models for which 0.1 The carbon deflagration model W7 This model (Nomoto et al. 1984) has been seen for a long time as the standard explosion model of SNIa. this is the case of Chandrasekhar models. Multidimensional simulations are presently conducted to understand these processes. The explosion synthesis of ∼ 0. is due to compressive heating from accretion. iii) the production of substantial amounts of intermediate-mass elements at expansion velocities of about 10 000 km s−1 near the maximum brightness of SNIa explosions. Because this work is a pedagogical introduction to the use of SNIa as cosmological probes.2 Delayed detonation models Since 1984. we only mention two Chandrasekhar mass models (groups of Nomoto and Woosley) that can account for the basic features of the so-called standard SNIa. i. in the core of a WD composed of C + O. The second important question is once ignited.2. All these features are very important to explain observed optical displays of SNIa: spectra and light curves. does the flame propagate supersonically by detonation (new fuel heated by shock compression ) ? or subsonically by deflagration (new fuel heated by conduction) ?.

2. the various spectral features can be identified as those of F e. S.3. ii) For heterogeneity Nomoto et al. the WD might undergo a pulsating detonation. Moreover. Spectra have been calculated for various models and can be used for comparisons with observations. However. by comparisons of predictive nucleosynthesis with solar isotopic ratio. (1996). DD4. in particular: progenitor evolution. effective propagation. Thus the full details of the combustion are quite uncertain. `a-priori. we briefly present some of the observed optical spectra and optical light curves of the so-called SNIa. type of the burning front (deflagration. we also give some considerations on γ-rays from SNIa. If the flame velocity is much smaller than in W7.which is still uncertain. speed of the burning front. Mg and O. Ca. constraints on these still uncertain parameters in models such as W7. For instance: i) Nugent et al. (1997) present observed optical spectra of SNIa (SN 1994D. detonation. One recalls that SNIa have not a thick H-rich envelope so that the elements which are synthesized during their explosions must be observed in their spectra. 2. synthetic spectra for the subChandrasekhar mass models are less satisfatory. As already noted above. Moreover. both at different stages and their transitions). WDD2 can be provided by comparisons of synthetical spectra and light curves with observations. the deflagration might also induce a detonation.1 On optical spectra of SNIa Optical spectra of SNIa are generally homogeneous but one can also observe some important variations. Si. SN 1990N) about one week before maximum brightness which show different features from those of SN1991T observed at the same 6 . a comparison between synthetic spectra and observations can be a powerful diagnostic of dynamics and nucleosynthesis of SNIa models. ignition densities. we only review the observations of SNIa which are relevant for our cosmological problem which can be formulated by the following question: Can one consider SNIa as a well-defined class of supernovae in order to use them as standard candles ? Therefore. for instance model DD4 of Woosley & Weaver (1994) or model WDD2 of Nomoto et al.3 On the observations of SNIa Here. Therefore. SN 1994A (20 days after the explosion) and synthetic spectra of the W7 model for the same epochs after the explosions. (1997) present agreement between observed spectra of SN 1992A (23 days after the explosion).

the flux at the maximum of the light curve for the 847 keV line and 1238 keV line -see for instance many γ-rays papers in Ruiz-Lapente et al. One must calculate.59 MeV including 1.2 × 10−4γ cm−2 s−1 was reported leading to an upper limit of M56 (3σ) < 0.2 On γ-rays from SNIa SNIa synthesise also radioactive nuclei such as 56 Ni.epoch. Let us also recall that SN 1991 T was quoted as a peculiar SNIa. the date.3.3 ± 2. Pinto et al.3.6 M⊙ ) is very important for the physics of the optical light curve of the SNIa. During each decay of 56 Co a number of γ-rays lines are produced. Let us note also that SN1991T is a spectroscopically peculiar SNIa. 2.4  D 2 M⊙ . of the 847 keV line has been reported. 57 Ni. • iii) An interesting test for the nature of SNIa explosion models has been found and published just after Cargese 2000. by the instrument OSSE/CGRO. 3 Mpc Later. A 3σ-upper limit F ≤ 4. for every SNIa model. from the point of view of optical spectroscopy (see above 2. for a certain time after the explosion the hard electromagnetic spectrum is dominated by γ-rays from the decays of 56 Co. As we will wee below. • ii) SN1991T (in NGC 4527) was observed by CGRO (Compton Gamma Ray Observatory) on day 66 after the explosion. An upper limit of: F847 ≤ 2.1). But in fact. 44 T i.1 × 10−5 γ cm−2 s−1 . the γ-ray light curve. From the observational point of view: • i) SN 1986G (in NGC 5128) was observed by SMM (Solar Maximum Mission). the amount of 56 Ni (∼ 0. (2000) present results of X and γ-ray transport calculations of Mch and sub-Mch explo7 . no detection.02 MeV from e+ − e− annihilations.5 × 10−5 γ cm−2 s−1 has been given only (see the γ-ray papers in Ruiz-Lapente et al. The average energy is about 3. 1997 and references therein). After 56 Co decays. SN 1986 G was estimated as an under luminous SNIa (Phillips 1993). The most prominent lines are the 847 keV line and the 1238 keV line. During the same date. 1997). other unstable nuclei become also important both for the energy budget and for producing observable hard emissions. The species 56 Ni is certainly the most abundant radioactivity produced in the explosion of a SNIa. A detection by the instrument COMPTEL/CGRO of the 847 keV line has been reported with a flux of F847 = 5. one must consider the following decay chain: 56 9 days 56 Ni −→ Co 112 days 56 −→ F e + e+ And.

all SNIa were identical explosions with identical light curve shapes and peak luminosities (Woosley & Weaver 1986.sion models for SNIa and they have shown that X-ray and γ-ray spectral evolution of both models are very different. see the upper plot of figure 2 in section 4. This was also supported by observations: for instance SN 1980N and SN 1981D in the galaxy NGC 1316 exhibited almost identical brightness and identical shapes of their light curves.3. the optical light curve shapes of SNIa were generally supposed to be homogeneous i. in light curve widths etc. despite all these differences. recent work on large samples of SNIa with high quality data has focused attention on examples of differences within the SNIa class: in particular there are observed deviations in luminosity at maximum light. several groups of theorists -experts in supernovaehave done a lot of work on SNIa in the early times and on their possible evolution until now -see most of the papers written in the proceedings edited by Niemeyer & Truran (2000) and those edited by Holt & Zhang (2000).. Anyway.e. The Phillips Relation For a long time. in Bmagnitudes. let us only notice that a first true detection of the 847 keV line of a true typical SNIa by INTEGRAL may help to better understand the SNIa models and the optical light curves of SNIa (see below 2. it is this emprical brightness decline relation (WLR) which allows the use of SNIa as calibrated candles in cosmology -see section 4Because SNIa are certainly more complex that can be described as a singleparameter supernova family. Phillips (1993) quantified the photometric differences among a set of nine well-observed SNIa using the following parameter ∆m15 (B) which measures the total drop. Then. in colors. from maximum to t = 15 days after B maximum.3 On optical light curves of SNIa. In effect. · The presence of surface 56 Ni in sub-Mch SNIa would make them very bright emitters of iron peak K-shell emission visible for several hundred days after explosion. It was this uniformity in the light curves of SNIa which had led to their primary use as standard candles in cosmology. This Phillips relation is also called the width luminosity relation (WLR). in particular. they show that: · The γ-ray light curves in sub-Mch models are much brighter and peak much earlier than in Mch SNIa. the Phillips relation: the brightest supernovae have the broadest light curve peaks. The 56 Ni γ-ray line emission (at 847 keV) from a bright sub-Mch explosion at 15 Mpc would be just at the limit for detection by the ESA satellite INTEGRAL (International Gamma Ray Astrophysics Laboratory) which will be launched in 2001. see also our table 1). the discovery of some regularities in the light curve data has also emerged. In particular.. However. 2. Finally.3). K-shell emission from a bright sub-Mch located near Virgo would be just above the limit for detection by the XMM observatory. But let us briefly recall how we can explain the present SNIa light curves and the 8 .3. In particular.

Anyway. Their decline is essentially due to the increasing transparency of the ejecta to γ-rays and to the decreasing input of radioactivity. 9 . the light curve models of SNIa depends on.Phillips relation or WLR. in particular. • ii) The adiabatic conversion of internal energy to kinetic energy of expansion. The theoretical peaks of the light curve is at about 15-20 days after the explosion. A priori. we will show how cosmologists use the Phillips relation to calibrate SNIa and use them as calibrated candles. Woosley & Weaver (in Petscheck. • 3) The small differences in initial conditions which might arise from evolutionary effects between z ∼ 0 and z ∼ 1 are unlikely to affect the SN cosmology results. To conclude this section 2. • iii) The escape of internal energy as the observed light curve. The light curve shape depends mainly on the diffusion time: τD ∼  κM 1/2 vesc c where κ is the opacity. Nomoto et al. vesc ∼ (E/M)1/2 . we will recall the cosmological background needed to use any standard candle for studying the geometry of the Universe -and in section 4. (1997) and more recently Pinto & Eastman (2000a. the empirical Phillips relation may be due to the radiation transport in SNIa and to atomic physics and not to hydrodynamics of the SNIa explosion (if the results of Pinto & Eastman are confirmed). In theoretical models of light curves. See. in the ignition time and nature of the burning front. 4) the opacity κ. four parameters: 1) the total mass M. in section 3. The light curves are also powered by the radioactive decay chain: 56 Ni →56 Co →56 F e. 2000c). 2000b. the question can be the following one : how these four (or five) parameters collapse to one for leading to the WLR ? Finally. the theoretical light curve of a SNIa is essentially determined by three effects: • i) The deposition of energy from radioactive decays. the explosion energy goes into the kinetic energy of expansion E. one must only mention again the work of Pinto & Eastman (2000c) which shows that: • 1) The WLR is a consequence of the radiation transport in SNIa. Arnett (1996). in other words. Therefore. 1990). 2) the explosion energy E. Or. 3) the 56 Ni mass or M56 . one can say that although there are uncertainties in the nature and evolution of SNIa progenitors. at least. • 2) The main parameter is the mass of radioactive 56 Ni produced in the explosion and the rate of change of γ-ray escape.

flat and closed.e. uµ being respectively the energy density. ds Therefore.Walker metric: ds2 = gµν dxµ dxν = dt2 − R2 h i dr 2 2 2 2 2 + r (dθ + sin θ dφ 1 − kr 2 (2) where R(t) is the cosmic scale factor. Equation (3) says us that three competing terms drive the expansion: a term of energy 10 . called the Hubble parameter: H2 ≡ R ˙ 2 R = k 8πGρ Λ + − 2 3 3 R (3) • An equation for the acceleration: ¨ R 4πG Λ =− (ρ + 3p) + R 3 3 (4) where R˙ = dR/dt and k is the curvature constant which equals respectively to -1. one must recall some theoretical background.+1 for a universe which is respectively open.1 The (Ωm . the Einstein field equations (1) simplify in: • The Friedman-Lemaitre equation for the expansion rate H. 3. the pressure and the 4velocity: dxµ uµ = . Tµν has the perfect fluid form: Tµν = pgµν + (ρ + p)uµ uν . Λ is the cosmological constant and Tµν is the energy.momentum tensor such as ∇ν T µν = 0.0.3 Cosmological Background In order to understand all the cosmological implications of observations of SNIa at high redshift. For a homogeneous and isotropic space-time. k is the curvature constant.ΩΛ )-plane The cosmological term -i. with ρ. gµν is the Robertson.was introduced by Einstein when he applied General Relativity to cosmology: Gµν = 8πGTµν + Λgµν (1) where gµν is the metric. the famous Λ term. p. Gµν is the Einstein tensor.

(matter and radiation). a term of dark energy or cosmological constant. Moreover signatures of ΩΛ are more subtle than signatures of Ho and ΩM . ωM = 0 for the cold matter. ωΛ = −1 for the cosmological constant. It is convenient to assign symbols to the various fractional constributions of the energy density at the present epoch such as: ΩM = Λ k 8πGρo . the Friedmann equation (3) can be rewritten:  da 2 dτ 1 = 1 + ΩM ( − 1) + ΩΛ (a2 − 1) a (8) ΩM and ΩΛ . Remarks: i) if one introduces the equation of state: p = ωρ. and a term of geometry or curvature term. the attempts to measure ΩΛ are infrequent and modest in scope. for this last case: −1 < ωQ < 0. at least from an observational point of view. 11 . are constant that parametrize the past (or future) evolution in terms of quantities of the present epoch. Then equation (3) can be written: Ωm + ΩΛ + Ωk = 1 (6) and the astronomer’s cosmological constant problem can be formulated by the following simple observational question: Is a non-zero ΩΛ required to achieve consistency in equation (6) ? In comparison with Ho and ΩM . not only consider a static. a priori. uniform vacuum density (or cosmological constant) but also a dynamical form of evolving . Anyway. here. inhomogeneous dark energy (or quintessence). let us recall that ωR = 1/3 for the radiation. If: a≡ 1 R = 1+z Ro (7) is the expansion factor relative to the present epoch and if τ = Ho t is a dimensionless time variable (i. ΩΛ = and Ωk = 2 2 3Ho 3Ho Ro Ho (5) where the index o refers to the present epoch. time in units of the measured Hubble time 1/Ho ).e. ii) Let us notice that one must. studying possible variations in equation (6) having dominant versus negligible ΩΛ -term is quite challenging ! Let us consider now the expansion dynamics in non-zero cosmological constant models.

ΩΛ )-plane.e. Of course. ΩΛ ) one gets different expansion histories. the most direct and theory independent way to measure Λ would be to actually determine the value of the scale factor a as a function of time. the expansion dynamics can be given by the system:  a˙ 2 a 1 1 = ΩM ( )3 + (1 − ΩM − ΩΛ )( )2 + ΩΛ a a a ¨ ΩM 1 3 = ΩΛ − ( ) a 2 a (11) (12) For different values of (ΩM . They are all related by simple z-factors (see for instance Weinberg 1972. but their analysis will be given in this (ΩM . magnitude. 12 . 3. 2 (10) Finally. i. of some class of objects -called standard candles.2. Peebles 1993. one will try to disentangle the effects of matter. Before presenting and discussing the recent results concerning type Ia-supernovae taken as standard candles. Here. ΩΛ )-plane which displays various possible regimes -see figure 1 of the (ΩM .1 On the luminosity distance DL In cosmology. 3. the results of high zsupernova observations will be first represented by the Hubble diagram.2 The luminosity distance DL The Hubble diagram is a graphic representation of the luminosity a function of their redshift. we recall some basic facts.Equivalently. it is very difficult ! However with sufficiently precise information about the dependence of a distance-measure on z. ΩΛ )-plane where it can be confronted to other major recent cosmological observational results. cosmological constant and spatial curvature. But. several different distance measurements are in use. we only give the figure of the (ΩM . Of course. one can also parametrize the evolution with ΩM and the deceleration parameter qo :  RR ¨ qo = − R˙ 2 o (9) which can also be written as: qo = ΩM − ΩΛ .

ΩM 1 | 3 - 2 | -3 No Big Bang Closed Universe -2 accelerating universe 2 qo = 0 @ R @ 1 . ΩΛ )-plane. the luminosity-distance DL is related to the transversecomoving distance Dm through: DL = (1 + z)Dm 13 (15) . ΩM . The one which is relevant for our study is the luminosity.Flat Universe decelerating universe -1 expands forever recollapses @ eventually @ R 0 - ΩΛ -0 h 6 Λ=0 -1 - - Open Universe | 1 0 | 2 3 Fig.distance: DL =  L 1/2 4πF (13) where L is the intrinsic luminosity of the source and F . the observed flux. For Friedmann-Lemaitre models. which depends parametrically on ΩM and ΩΛ defined by equations (5). : (ΩM . 1. one can show that: DL (z) = c dL(z. ΩΛ ) = DH dL Ho (14) where DH ≡ c/Ho is the Hubble distance and dL is a known dimensionless function of z. In effect. Hoggs 1999).

In other words and as we will see. Ho ) and dL = dL (z. if Ωk > 0 (16) Zz (17) • D m = Dc = DH 0 dz ′ if Ωk = 0 E(z ′ ) with E(z) = Ωm (1 + z)3 + Ωk (1 + z)2 + ΩΛ ]1/2 1/2 • Dm = DH Ω−1/2 sin[Ωk Dc /DH ]. from equations (6) and (14-18) we check that. ΩM .3 On the magnitude redshift relation The apparent magnitude m of an object is related to its absolute magnitude M through the distance modulus by: m − M = 5log10 DL + 25. 14 . ΩΛ ) (19) 3. ΩΛ . • ii) It is the comparison of the theoretical expectation (21) with data which can lead to constraints on the parameters ΩM and ΩΛ . ΩM . the likelihoods for cosmological parameters ΩM and ΩΛ will be determined by minizing the χ2 statistic between the measured and predicted distances/magnitudes of SNIa taken as standard candles. ΩM . ΩΛ ) + M (21) Let us only note here that: • i)M ≡ M − 5log Ho − 25 is a non-important fit parameter. from equation (14-18) and (20) we obtain a relation between the apparent magnitude m and z with ΩΛ and ΩM as parameters: m = 5log10 dL (z. Mpc (20) Therefore.with: 1/2 • Dm = DH Ω−1/2 sinh[Ωk Dc /DH ]. if Ωk < 0. (18) Therefore. for different possible values of Ωk : DL = DL (z.

with the Hubble Space Telescope (HST). In this section. we briefly review the strategy described by the SCP for its early campaigns. we briefly presented SNIa and emphasized that there are large uncertainties in the theoretical models of these explosive events as well as on the nature of their progenitors. Here.developed a strategy that garantee the discovery of many SNIa on a certain date. convincing evidence has been found for a correlation between light curve shape and luminosity of nearby SNIa -brighter implies broader. 1999 and references therein). they observed some 50-100 high galactic latitude fields -each containing about 1000 high-z galaxies. 1998.1 The observations Both groups -the SCP and the HZT. In Section (2). (1998 and all references therein). at the Cerro-Tololo 4m telescope in Chile with a Tyson and Bernstein’s wide-field-camera.which has been quantified by Phillips (1993).1. Isaac Newton Telescope (INT) and -for the hightest z SNIa. a discussion of these results and their cosmological implications. Perlmutter (see S. Schmidt (see B. • ii) The High z-Supernovae Search Team (HZT) which is led by B. (1997. 15 . 4. we briefly present the cosmological use of SNIa at high redshift with: the observations.4 Observations of SNIa at high redshift There are two major teams investigating high-z SNIa: • i) The Supernova Cosmology Project (SCP) which is led by S. the results. Schmidt et al. We have also seen that although they cannot be considered as perfect standard candles. 4.1 On the strategy Just after a new moon. They returned three weeks later to observe the same two nights. About two dozen SNIa were discovered in the image of the ten thousands of galaxies which were still brightening since the risetime of a SNIa is longer than 3 weeks. They observed the supernovae with spectroscopy at maximum light at the Keck-Telescope and with photometry during the following two months at the Cerro-Tololo International Observatory (CITO). Both groups have published almost identical results. Perlmutter et al.

In principle. 16 . in the expanding atmosphere. This correlation. its spectrum changes showing on each day which elements. (2) shows how the stretch correction aligns both the light curve width and the peak magnitude for the nearby Hamuy supernovae. one has evidence that SNIa have evolved between the epoch z and now. 1998).1. their spectra must give a constraint on high-z SNIa: they must show all of the same features on the same day after the explosion as nearby SNIa.1.are plotted in the Hubble diagram as a function of their host galaxy redshift. often called Phillips relation or luminosity width relation (LWR) has been refined and described through several versions of a one-parameter brightness decline relation: i) ∆m(15) (Hamuy et al. (1998) computed a peak-magnitude in the B-band corrected for galaxy extinction and the stretch parameter that stretches the time axis of a template SNIa (see 4.4. For each supernova: • i) The final image of the host galaxy alone is substracted from the many images of the given SNIa spanning its light curve.2 On spectra With the Keck (10m-telescope): i) they confirmed the nature of a large number of candidate supernovae.3 On light curves As already noted. (1998).3) to match the observed light curve. iii) Stretch parameter (Perlmutter et al.4 On the analysis One can summarize the three-steps analysis of the 42 high-z SNIa presented in Perlmutter et al. all of the SN magnitudes -corrected for the stretch-lumisosity relation. • ii) Perlmutter et al. ii) they searched for peculiarities in the spectra that might indicate evolution of SNIa with redshift. Fig. 4. 1996). in Section 2. 1996): ii) multicolor light curve shape (MLCS) (Riess et al. 4. nearby SNIa show a relationship between their peak luminosity and the timescale of their light curve: brighter implies broader. In effect. • iii) Then. as a supernova brightens and fades.1. Moreover. or else. are passing through the photosphere.1. there exists a simple linear relation between the absolute magnitude and the stretch parameter.

2. ΩM . : The upper plot shows nearby SNIa light curves which exhibit LWR. 4. after the stretch parameter correction. (1997) Fig. et al. measured in the upper plot. the lower plot shows the nearby SNIa light curves.2 The results and discussion One must compare the redshift dependence of observed m with the theoretical expectation given above in Section 3: m = 5 log dL (z. ΩΛ ) + M 17 (22) .B Band -20 as measured MB – 5 log(h/65) -19 -18 -17 -16 Calan/Tololo SNe Ia -15 -20 0 20 40 60 days -20 light-curve timescale “stretch-factor” corrected MB – 5 log(h/65) -19 -18 -17 -16 -15 -20 0 20 40 60 days Kim.

(0.5 0. 1996) (a) 14 1. (ΩM . (1. The dashed curves are for flat cosmological models (ΩM +ΩΛ = 1): (ΩM . (3).2 0.2. The solid curves are the theoretical mef B (z) for a range of cosmological models with Λ = 0.0 redshift z Fig.0) in middle.5 ) 0.8 1.5. : From Supernova Cosmology Project (SCP): Hubble diagram.0) on bottom. ΩΛ ) = 18 .0. (0.5) (2. ΩΛ ) = (0. 0) ( 1.on a linear redshift scale to display details at f high z.1.5 1.J. (0.4 0.4 above). 1) (0.0 -0.The effective magnitude versus redshift can be fitted to various cosmologies.1 On the Hubble diagram Flat 24 Λ=0 (ΩΜ. 3. In Fig.5. ΩΛ) = 6 4 2 0 -2 -4 -6 0. 0.75. 0 ) (1. 0) effective mB 22 Supernova Cosmology Project 20 18 16 Calan/Tololo (Hamuy et al. A. 1 ) (0.–0.0 (b) (0.0 -1.72) 0) 0. 0) (1. 4.25 ) 0) (c) 0.ΩΛ) = ( 0.6 0.5) (0.5 -1. 0) on top: (1. one sees the Hubble diagram for 42 high-z SNIa from the SCP and 18 SNIa from the Cal`an-Tololo Supernovae Survey -after correcting both sets for the LWR (see 4. (2. in particular: • i) to the flat universes (ΩM + ΩΛ = 1) • ii) to the Λ = 0-universes.0 standard deviation mag residual (ΩΜ .

5) in middle. (0. (1.2 On the (ΩM . The best fit (Perlmutter et al. in their last Hubble diagram Kim (2000).f.(0. 4. (1998) 3 No Big Bang 99% % 95 % 0 9 42 Supernovae 2 (cosmological constant) vacuum energy density 68% ing rat e l e acc ating r ele c e d 1 expands forever lly recollapses eventua 0 cl os e fla d op t en -1 1 2 mass density Two groups results agree: c. Note that. 1) on top. more analysis and more SNIa at z > 1 are needed.28. The middle panel of Fig. (3) shows the magnitude residuals between data and models from the best fit flat cosmology: (ΩM .5. 1998) for a flat Universe (ΩM + ΩΛ = 1): ΩM ∼ 0. ΩΛ ) = (0. ΩΛ ∼ 0. ΩΛ )-plane Supernova Cosmology Project Perlmutter et al. But. (1998) 3 h's ut G of eory n tio th ic n" ed io pr flat "in 0 Fig.5) on bottom. 0.5. : From Supernova Cosmology Project (SCP): the (ΩM .72). The bottom panel shows the standard deviations. Perlmutter (2000) extend it beyond z = 1 by estimating the magnitude of SN1999eq (Albinoni).0. 19 . ΩΛ )-plane. 4.72.2. Riess et al.28.0.

1 If one assumes a flat Universe (ΩM + ΩΛ = 1). ΩΛ )-plane for the preliminary analysis of the 42 SNIa fit presented in Perlmutter et al. flat or closed.5 at 95 % confidence level).04 · Gravitational lensing by clumped mass: δM < 0.2 ± 0.04 (syst. (1998). From these figures. one can say that the data are strongly inconsistent with the Λ = 0. one must try to discuss the believability of these results. Note that the spatial curvature of the universe -open. flat universe model.8ΩM − 0. Here.03 • ii) Additional sources of systematic uncertainties: · Extinction by gray dust 20 .05 ΩfMlat = 0. (4) shows the best fit confidence regions in the (ΩM .3 Discussion of the results The current SNIa data set already has statistical uncertainties that are only a factor of two larger than the identified systematic uncertainties. But before interpreting the results of the analysis of the SNIa data. (1998). the supernova data require a non-zero cosmological constant at a high degree of significance (ΩΛ ≥ 0. 4. The analysis of Perlmutter et al. If the simplest inflationary theories are correct and the universe is spatially flat.05 · Galaxy extinction: δM < 0. For a spatially flat universe. let us only list the main indentified sources of systematic uncertainty and only mention some additional sources of systematic uncertainty (from Perlmutter 2000): • i) Identified sources of systematic error and their estimated contribution to the uncertainty of the above measurements: · Malmquist bias: δM ∼ 0. (1998) suggested a universe with: 0.) More recent data of both groups (SCP and HST) and other independent likelihood analysis of the SNIa data provide robust constraints on ΩM and ΩΛ consistent with those derived by Perlmutter et al. the data imply: +0.28+0.04 · κ-correction and cross filter calibration: δM < 0.09 −0. indicated by the nearhorizontal solid line.06 · Extinction by extragalactic dust:δM ∼ 0.Fig.6ΩΛ ∼ −0. then the supernova data imply that there is a significant positive cosmological constant.08 (1σstat )−0.03 · Non-SNIa contamination: δ < not deterninative of the future of the universe’s expansion.

2000b. a 2-m satellite telescope proposed by Perlmutter (2000) 2 . 2 see http://snap. The distribution of these stellar properties may evolve in a given galaxy and over a set of galaxies. by Straumann (1999). Ruiz-Lapente et al. Such supernova searches. for instance. the believability of the high-z SNIa results turn on the reliability of SNIa as one-parameter standard candles. Weinberg (2000) and references therein. (1997) . due to the lack of a good theoretical understanding of the Phillips relation (what is the physical parameter ?). Note also that a very optimistic estimate of the limiting systematic uncertainty of one percent is expected for such supernovae searches.· Uncorrected evolution: SNIa behaviour may depend on properties of its progenitor star or binary-star system. For a while. an accelerating universe and strongly rule out ΩΛ = 0 universes. The one-parameter is the rate decline of the light curve (the brighter implies the broader) -see subsection 2. The classic discussion of the physics of the cosmological constant is by Weinberg (1989). More recent works are discussed. December 1999: my personal answer to the question of whether the Universe is accelerating is probably yes. Carroll (2000). the supernova experts were not completely convinced by the high-z SNIa observations and results.3. the confidence levels in the (ΩM − ΩΛ )-plane -which are consistent for both groups (SCP and HZT). 2000c) -the results of (SCP) and (HZT) appear to supernova experts almost quite convincing. and of course the papers of Pinto & Eastman (2000a.4 Implications of the results. 4. in particular. extended to greater z. The dark energy Therefore. After several meetings and a lot of theoretical papers -see. For a detailed discussion of the present state of all the systematic uncertainties we refer the reader to the paper of Riess (2000) and references therein. In other words.lbl. Let us only recall the conclusion of Wheeler (summary talk of Cosmic Explosions. Anyway. There are a number of reviews on the various aspects of the cosmological constant. Holt & Zhang (2000).gov 21 . the current results of high-z SNIa observations suggest that the expansion of the Universe is accelerating indicating the existence of a cosmological constant or dark energy.. can make a much precise determination of the luminosity distance as a function of z. dL (z).. if one accepts the high-z SNIa results. Let us only notice that the control of statistical and systematic uncertainties is one of the main goals of the Supernovae Acceleration Probe (SNAP). Niemayer & Truran (2000).3 above. My answer to the query of do we know for sure is not yet !.favor a positive cosmological constant.

1. Spergel & Pen 1996. One possibility is that the dark energy consists of vacuum energy or cosmological constant. August 2000): There are dumb trackers (Steinhardt et al.3. a smooth energy component with negative pressure and negative equation of state. i. Telescopes may play a larger role than accelerators in determining the nature of the dark energy. 2000b.2. For the future -and as already noted in the subsection 4. Anyway.1 < z < 1. the pressure is negative and the equation of state is a function of z. In this context. For this case. ρ An other possibility is quintessence a time-evolving. it is expected that about 1000-2000 SNIa at greater z (0. 1999). for instance. Vancouver.of the same order of magnitude as the present mass density of the universe ? The challenge for cosmology and for fundamental physics is to determine the nature of the dark energy. spatially inhomogeneous energy component.e. In this case -as already noted above in the remarks of the subsection 3.In particular Weinberg (2000) formulates the two cosmological constant problems as: • i) Why the vacuum energy density ρv (or ΩΛ ) is not very large ? • ii) Why ρv (or ΩΛ ) is not only small but also -as high-z SNIa results seem indicate. by observing SNIa at z about 1. a precise measurement of ω today and its time variation could provide important information about the dynamical properties and the nature of dark energy. Let us only notice here that: • i) these models of dark energy are based on nearly-massless scalar fields slowly rolling down a potential. models based on networks of tangled cosmic strings for which the equation of state is: ωstring = −1/3 or based on walls for which ωwall = −2/3 (Vilenkin 1984. smart trackers (Armandariz-Picon et al. 1998). SCP and HZT have found strong evidence for dark energy i. Battye et al. Bean & Magueijo 2000).7) will be detected. at least for part of its history.the equation of state of the universe is ωΛ ≡ p = −1. As Steinhardt said (in the introductory talk of the Workshop on String Cosmology. 2000a. ωQ (z) is such that: −1 < ωQ (z) < 0. most of them are called tracker models of quintessence for which the scalar field energy density can parallel that of matter or radiation.the spatial mission SNAP is being planned. SNAP has also very ambitious goals: it is expected that SNAP will test the 22 . • ii) there are other models of dark energy besides those based on nearlymassless scalar fields.e.

will better determine the equation of state of the universe and its time variation i. What are SNIa and SNIa progenitors ? Are they single or double degenerate stars ? Do they fit Chandrasekhar or sub-Chandrasekhar mass models at explosion ? Is the mass of radioactive 56 Ni produced in the explosion really the main parameter underlying the Phillips relation ? If so. 2000) show that the determination of the equation of state of the Universe by high-z SNIa searches -by using luminosity distance.P. this will proved to be a remarkable discovery. Acknowledgements The authors gratefully acknowledge Stan Woosley for helpful suggestions. because of the preexisting evidence for dark energy that led to the prediction of accelerated expansion. has been conducted under the auspices of the Dr Tomalla Foundation and Swiss National Science Foundation. If verified. 2000a. ω(z). References Armandariz-Picon C.are strongly limited.D. et al. many questions without definitive answers remain. However. astro-ph/0006373 Arnett W. we presented the work done by two separate research teams on observations of high-z SNIa (Section 4) and their interpretation (Section 5) that the expansion of the Universe is accelerating. Press 23 . 1996 in Nucleosynthesis and supernovae. astro-ph/0004134 Armandariz-Picon C. like SNAP. et al. cosmologists are inclined to believe the SNIa results.nature of the dark energy. Part of the work of D. provide large homogeneous samples of very distant SNIa and help determine several key cosmological parameters with an accuracy exceeding that of planned CMB observations ? Meanwhile. 2000b. how important are SNIa evolutionary corrections ? Can systematic errors be negligible and can they be converted into statistical errors ? Can a satellite. Cambridge Univ.e. 5 Conclusions After a brief review of type Ia supernovae properties (Section 2) and of cosmological background (section 3). But very recent studies (for instance Maor et al.

Riess A. Elsevier 24 . Phys. 1993 BAAS 182. 2000. 53.. 419 Vilenkin A. et al. 2000. astro-ph/0005344 Maor Kim A.. et al. astro-ph/9905049 Fillipenko A. 2000.lbl. 2000.. 2000. 1997. 20. 1994 in Les Houches. 205 Woosley S. et al. 46 SNAP (Supernova Acceleration Probe). et al. 1984. Lett. 1999 Eur. 1993 in Principles of Physical Cosmology. Kluwer eds Schmidt B. 1997 ApJ 485. astro-ph/9812313 Straumann N. 744 Pinto P. Press Nomoto K. 507. 61. in Type Ia Supernovae: Theory and Cosmology Cambridge Univ. astro-ph/0008057 Hoggs D. 2398 Fillipenko A. 1999. 51 Perlmutter Spergel D. J. 1999 ARAA 36. Phys.. 756 Pinto P. Rev. 1998 ApJ. 2000 in Cosmic Explosions. 1984. et al. astro-ph/9912403 Woosley S. Eastman 2000c. astro-ph/0008330 Riess A.. Eastman 2000b ApJ 530. 588 Riess A. Princeton Univ.. Wiley Eds Weinberg S. 1999. astro-ph/0004075 Hamuy M. Truran J. et al 1997 ApJ 483. Press Perlmutter S. Magueijo J. Weaver T. 1997 in Thermonuclear supernovae. 1989. et al. 1999 ApJ 517. ApJ 473. Weaver T. Eastman 2000a ApJ 530.. 17 Carroll S. 2907 Pinto P. et Petschek A.Battye R. astro-ph/0007297 Niemeyer J. 2000. et al. ApJ 286. 1998. Rev. Livio M..lbl. et al. 644 Nomoto K. 1972 in Gravitation and Cosmology. Zhang W. session LIV. et al. et al. et al. 2000. 1016 Weinberg S. astro-ph/9905116 Holt S.. astro-ph/9908047 Bean R. http://snap. http://snap. http://www-supernova. Phys. Pen U. 1990 in Supernovae. AIP in press Kim A. Springer-Verlag Phillips M.. 2000. et al. AJ 112. L67 Steinhardt P. 1997 ApJ 491. 1997. Riess A. astro-/ph/9706007 Nugent et al. Bludman et al. 812 Peebles J. et al. 1998 Nature 391. Mod. 365 Perlmutter S. 2000. 1986 ARAA 24. 1999. 1 Weinberg S. 1999. 2000. astro-ph/0006171 Pinto P. 565 Perlmutter S. 1996. http://snap. astro-ph/0007199 Branch D. 2000. astro-ph/0005265 Wheeler S.. astro-ph/0005229 Ruiz-Lapente R. S.